INDEX

Explanations

references to the reader or audience

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 22-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L22R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.22.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 neither

-0.16

not

-0.15

 only

-0.14

IXEL

-0.14

 done

-0.14

 shouldn

-0.14

 cannot

-0.14

 Î´Îµ

-0.14

UBL

-0.14

 never

-0.13

POSITIVE LOGITS

 exactly

0.20

Ú¯ÙĪ

0.16

èĳ

0.14

ãĥ¼ãĤ¹ãĥĪ

0.14

(åľŁ

0.14

éı

0.14

 pÅĻesnÄĽ

0.14

ä¹ħä¹ħ

0.14

 Intelli

0.14

adow

0.13

Activations Density 0.075%

references to the reader or audience

No Comments

No Known Activations